INTRO: Selection of the right trackform for an underground railway has been simplified by the development of a simulation model that accurately predicts the transmission of ground-borne sound and vibration from the rail surface to nearby buildings

BYLINE: Ulrik Danneskiold-Samsøe, Uffe Degn and José Luis Eguiguren*

BYLINE: * Ulrik Danneskiold-Samsøe is Marketing Director and Uffe Degn is Department Manager at Ødegaard & Danneskiold-Samsøe A/S in København. José Luis Eguiguren is Project Manager for Labein in Bilbao

PREDICTION OF the sound and vibration level in buildings close to railway tracks has always been difficult, because of inaccuracies in the standard models for transmission loss in the signals that propagate through the ground. However, a project to move underground an existing railway in Bilbao is amongst the first to benefit from a new modelling technique.

When predicting the levels of noise and vibration in buildings close to railway alignments, there are four main elements to address: wheel/rail contact, the track and tunnel design, the ground-borne transmission mechanism, and the architecture of the buildings affected.

In recent years, various methods have been developed to measure the roughness and out-of-roundness of wheels. This means that the definition and calculation of vibration and structure-borne noise levels caused by rough wheels and rails is no longer a problem - given the appropriate roughness data.

Similarly, new track designs that combine resilient components and blocking masses enable a better reduction of the vibration and structure-borne noise level. These designs can be applied to both surface and tunnelled alignments. Once again, their performance in reducing noise and vibration can be modelled, measured and predicted with considerable accuracy.

Propagation of vibration and structure-borne noise in buildings is a part of classical architectural acoustics, and various models already exist for the prediction of vibration and noise levels in different rooms. These predictions are based on the noise and vibration levels reaching the lower part of the structure, depending upon the distance from the railway and the properties of the intervening ground.

Until recently, the prediction of ground transmission losses has been the largest contributor to inaccurate predictions. But development of a new proprietary programme to predict these transmission losses, based on known geotechnical properties, has to a large extent eliminated this problem. The knowledge gap between the rail surface and the interior of nearby buildings can now be considered as closed.

The GROund Noise And Vibration simulation tool (ODS-GRONAV) has been developed in co-operation between Ødegaard & Danneskiold-Samsøe and Prof Steen Krenk, from the Department of Mechanical Engineering at the Technical University of Denmark. The main feature of the tool is a new transmission condition for elastic waves developed by Dr Krenk. The theory behind this condition is about to be submitted for scientific publication.

ODS-GRONAV is based on a 2D finite element model. This includes specially-developed elements providing properties similar to an infinite domain. The finite element technique allows precise calculation of wave propagation in complex ground with different layers, tunnel structures and building foundations.

The finite element model has to be limited in space to keep within a reasonable calculation capacity. This would normally cause problems with reflection of the propagating waves at the model boundaries. However, special transmitting boundary elements have been developed to solve this problem. Rayleigh, P and Swaves are all included.

Using the model, forces are applied at the track, and the vibration velocity level is calculated in the soil in front of the buildings. Table I lists the geotechnical parameters needed to calibrate the model for a specific location.

Going underground in Bilbao

An early application for ODS-GRONAV has been the design of a new tunnel in Bilbao. As part of the Bilbao RIA 2000 urban remodelling project, the existing surface railway is to be relocated underground through an historic part of the city centre. The investigation was undertaken for the City of Bilbao, as a joint project between ØDS and local technical consultancy Labein.

The project required the measurement of sound and vibration levels generated by the existing surface line and the prediction of the sound and vibration levels which would emanate from the future underground alignment.

Using the available geotechnical data, two models were developed - one for the surface alignment and one for the tunnel. The surface model was validated by measurement of the source strength and the actual vibration and structure-borne transmission loss between rail surface and the foundation of a nearby building, using impact testing methods.

Measurements were taken during the passage of a train on the existing surface track. Fig 1 shows a satisfactory fit between the predicted and measured figures recorded at a distance of 6m from the track.

Uncertainties in the prediction arose from the limited accuracy of the geological model of the subsurface. This was based upon spot tests made by drilling holes in the soil, but a heavy concrete layer proved very difficult to penetrate to reach the subsurface layers.

The second model was then used to predict the structure-borne sound level in the building after the tunnel had been completed.

At Location C the lower half of the tunnel was to be excavated through hard and heavy siltstone, compared to the much lighter and soft dry clay which supported the existing track. Fig 2 shows the result, with a significant improvement in structure-borne sound levels as a result of ’locking’ the tunnel into such a solid foundation layer.

The opposite situation was experienced at Location D, where the tunnel is to be built through the same dry clay that supports the current track. It can be seen in Fig 3 that the improvement is much less than that expected at Location C.

Scope for noise reduction

The transmission path for vibration and structure-borne noise provides many opportunities for reduction measures ranging from the wheel/rail interface to individual rooms in the affected buildings.

Firstly, the sensitivity of the rooms has a major impact on the level of measures that need to be taken. There is a huge difference in sensitivity between a motor car repair shop and a concert hall or theatre stage. Once the sensitivity level has been determined there are a number of options that can be pursued.

The greatest cost/benefit comes from addressing the wheel/rail interface. Smooth surfaces can be ensured both by good design and by proper maintenance. The impact from wheel flats or rail joints should be completely eliminated. Use of a modern resilient slab track is also beneficial. Although the costs are higher than for traditional ballasted track, life-cycle cost savings may be obtained from the use of a smaller tunnel diameter and reduced long-term maintenance requirements.

Finally, the selection of materials for the subsurface layer is important. A heavy, stiff subsurface is better than a light, soft one.

TABLE: Table I. Geotechnical parameters needed for ODS-GRONAV


Layer topology

Layer material type

Ground water level

Per subsurface layer


Poisson’s ratio


Loss factor

Pressure wave velocity

Shear wave velocity

Amplification factor

CAPTION: Fig 1. Comparison of predicted and measured frequency response for vibration level using the existing surface track at Location C in Bilbao

CAPTION: Fig 2 (above) demonstrates the large reduction in noise and vibration levels that can be expected when a tunnel is anchored in a hard and heavy subsurface. Fig 3 (right) shows that the difference is much less when the tunnel is located in a light and soft subsurface

CAPTION: LEFT:A snapshot of the predicted wave propagation pattern for the existing line as recorded at the arrowed location, which was subsequently validated by practical tests using an impact hammer (above left)

RIGHT: Ground conditions (above) and a snapshot of the predicted wave propagation pattern for the proposed new works, with a double-track tunnel for passenger trains and a single track bore to carry freight traffic